8/6/2012

1

Thoughts on Verification Your Professor Thoughts on Verification Your Professor May Not Have Mentioned May Not Have Mentioned

Michael ValenteMichael Valente

Thoughts on Verification Your Professor Thoughts on Verification Your Professor May Not Have Mentioned May Not Have Mentioned

Michael ValenteMichael Valente

Siemens Expert Siemens Expert Series Series

Aug 10, 2012Aug 10, 2012

OutlineOutline

• Introduction to the use of REM by audiologists

• Example of first-fit versus programmed

• Factors to be discussed:– REIG using average versus measured REUG

• Introduction to the use of REM by audiologists

• Example of first-fit versus programmed

• Factors to be discussed:– REIG using average versus measured REUGg g

– Correcting for bilateral fit

– Correcting for the number of channels of signal processing

– Average versus measured REDD and impact on REAR measures

g g

– Correcting for bilateral fit

– Correcting for the number of channels of signal processing

– Average versus measured REDD and impact on REAR measures

KirkwoodKirkwoodHear J (2010)Hear J (2010)

KirkwoodKirkwoodHear J (2010)Hear J (2010)

0%0% OccasionallyOccasionally <50%<50% 50%50% Most Most of of TimeTime

Almost Almost AlwaysAlways

AudiologistAudiologist 29.929.9 19.519.5 13.213.2 7.17.1 11.511.5 19.219.2

“How often do you do REM? “How often do you do REM?

gg

HISHIS 27.027.0 21.121.1 8.98.9 8.98.9 13.013.0 21.121.1

8/6/2012

2

Mueller and Mueller and PicouPicouHear J (2010)Hear J (2010)

Mueller and Mueller and PicouPicouHear J (2010)Hear J (2010)

Seldom or Seldom or NeverNever

SometimesSometimes 50%50% UsuallyUsually Almost Almost AlwaysAlways

Audiologist Audiologist 34%34% 18%18% 6%6% 12%12% 30%30%

“How often do you do REM? “How often do you do REM?

ggand HISand HIS

Typical FirstTypical First--Fit Fit

REIG With VerificationREIG With Verification

8/6/2012

3

AazhAazh et al 2012et al 2012“The Accuracy of Matching Target Insertion Gain with Open“The Accuracy of Matching Target Insertion Gain with Open--

Fit Hearing Aids”Fit Hearing Aids”

AazhAazh et al 2012et al 2012“The Accuracy of Matching Target Insertion Gain with Open“The Accuracy of Matching Target Insertion Gain with Open--

Fit Hearing Aids”Fit Hearing Aids”

• N = 51 fittings

• 71% of initial fits failed to be within 10 dB of NAL-NL1 at one or more 7 discrete frequencies between 250-4000 Hz.

• 10 dB is a rather liberal margin for a “acceptable” fit.

• After modification and verification using REM, 18% failed to be within 10 dB of target in one or more frequencies between 250-

• N = 51 fittings

• 71% of initial fits failed to be within 10 dB of NAL-NL1 at one or more 7 discrete frequencies between 250-4000 Hz.

• 10 dB is a rather liberal margin for a “acceptable” fit.

• After modification and verification using REM, 18% failed to be within 10 dB of target in one or more frequencies between 250-g q4000 Hz.

• Several studies (Swan and Gatehouse, 1995; Harrowven, 1998; Norman and James, 2000; Hawkins and Cook, 2003; Aarts and Caffee, 2005; Aazh and Moore, 2007) reported similar findings.

• However, the results from these and other studies as well as two national guidelines (ASHA, 1998; AAA, 2006) have failed to convince most audiologists to routinely use REM to verify the fitting of hearing aids.

g q4000 Hz.

• Several studies (Swan and Gatehouse, 1995; Harrowven, 1998; Norman and James, 2000; Hawkins and Cook, 2003; Aarts and Caffee, 2005; Aazh and Moore, 2007) reported similar findings.

• However, the results from these and other studies as well as two national guidelines (ASHA, 1998; AAA, 2006) have failed to convince most audiologists to routinely use REM to verify the fitting of hearing aids.

But, for the 20But, for the 20--30% who 30% who routinely perform REM’s, you routinely perform REM’s, you

But, for the 20But, for the 20--30% who 30% who routinely perform REM’s, you routinely perform REM’s, you

might not be aware of…..might not be aware of…..might not be aware of…..might not be aware of…..

Mueller and Picou (2010)Hear J 63(5)27-28,30,32

78% use REAR; 22% use REIG, while some use both.

Mueller and Picou (2010)Hear J 63(5)27-28,30,32

78% use REAR; 22% use REIG, while some use both.

91% of audiologists using REIG use measured REUG

91% of audiologists using REIG use measured REUG

8/6/2012

4

REIG Using REIG Using AverageAverage REUGREUG

Average REUG

REIG Using REIG Using MeasuredMeasuredREUGREUG

8/6/2012

5

Individual REUG

Valente et al 1991Valente et al 1991Valente et al 1991Valente et al 1991

N = 49 ears

-5

0

5

10

15

20

25

30

35

40

Gai

n (

dB

SP

L)

Average REAGMeasured REAGDifferenceLinear (Difference)

-10200 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000

Frequency (Hz)

If measured REUG @ any Hz is > than average REUG then > REAG is required to reach target REIG. If measured REUG @ any Hz is < than average REUG, then < REAG is required to reach target REIG.

8/6/2012

6

…..the Profession is shifting to REAR…..the Profession is shifting to REAR

1. Audiology is shifting from REIG to REAR because of the popularity of Verifit and Live Speech Mapping (LSM). Also, REAR measures intuitively make great sense.

2. REAR measures can be SPL-O-Gram or LSM “formats.”

3. In my opinion, SPL-O-Gram is a verification tool; LSM is a counselingtool (no standardized signal or input level; no validated target of measured SL).

4. SPL-O-Gram uses DSL v5 or NAL-NL1(2) targets for single (linear) or ( ) g g ( )multiple (nonlinear) input levels.

5. Both methods use the DR (in dB SPL) as the target. The DR, however, in both methods is typically predicted based on average transformationsafter entering the audiogram.

6. Whether one uses REIG or REAR there are errors many may not be aware of.

…..failure to correct for bilateral …..failure to correct for bilateral andandchannel summation channel summation

…..failure to correct for bilateral …..failure to correct for bilateral andandchannel summation channel summation

8/6/2012

7

65 dB SPL65 dB SPL

AmplifierAmplifier

An Analogy to Help Explain An Analogy to Help Explain Channel Channel SummationSummation

SLMSLM

Now, add 11 loudspeakers and keep the input level to each

l d k l

Now, add 11 loudspeakers and keep the input level to each

l d k lloudspeaker equalloudspeaker equal

65 dB SPL65 dB SPL

AmplifierAmplifier

An Analogy to Help Explain An Analogy to Help Explain Channel Channel SummationSummation

SLMSLM

8/6/2012

8

Entered audiogram for fitting bilateral Entered audiogram for fitting bilateral hearing aids with 20 channels of signal hearing aids with 20 channels of signal

processing.processing.

As a sidebar....notice the predicted LDL’s As a sidebar....notice the predicted LDL’s in dB HLin dB HL. . A A bit high wouldn’t bit high wouldn’t you say?you say?

REM default is REM default is monauralmonaural and and 1 channel,1 channel, but but audiologists typically do not “see” this.audiologists typically do not “see” this.

REIG65 for the entered audiogramREIG65 for the entered audiogram

8/6/2012

9

Now...I’m going to correct for bilateral summationNow...I’m going to correct for bilateral summationNow...I’m going to correct for bilateral summationNow...I’m going to correct for bilateral summation

Resulting REIG for bilateral summationResulting REIG for bilateral summationResulting REIG for bilateral summationResulting REIG for bilateral summation

DefaultDefault

Corrected Corrected

Now...we’re going to correct for the number of Now...we’re going to correct for the number of channelschannels

Now...we’re going to correct for the number of Now...we’re going to correct for the number of channelschannels

8/6/2012

10

Resulting REIG for binaural Resulting REIG for binaural andand channel summationchannel summationResulting REIG for binaural Resulting REIG for binaural andand channel summationchannel summation

Want to see that again?Want to see that again?Want to see that again?Want to see that again?

InitialInitial

Difference

BilateralBilateral

5 dB

Bilateral + Channel Bilateral + Channel Difference

250 5500 5 1000 72000 114000 66000 6

…..third error when doing …..third error when doing REAR MREAR M

…..third error when doing …..third error when doing REAR MREAR MREAR Measures REAR Measures REAR Measures REAR Measures

8/6/2012

11

Real Ear to Dial Difference (REDD)

Real Ear to Dial Difference (REDD)

When completing REAR measures you may When completing REAR measures you may not be aware….not be aware….

When completing REAR measures you may When completing REAR measures you may not be aware….not be aware….

When you enter the audiogram in dB HL, the REM software will convert the threshold (dB HL) to dB (SPL) by adding the average REDD from ANSI S3.6-1989 to calculate the predicted threshold in dB SPL:

When you enter the audiogram in dB HL, the REM software will convert the threshold (dB HL) to dB (SPL) by adding the average REDD from ANSI S3.6-1989 to calculate the predicted threshold in dB SPL:

250 19 3000 15.5

500 12 4000 13

1000 9 6000 13

2000 15 8000 14

250 19 3000 15.5

500 12 4000 13

1000 9 6000 13

2000 15 8000 14

Predicted Threshold

8/6/2012

12

20

25

30

35

40

RE

DD

(in

dB

))

500 Hz

1000 Hz

2000 Hz

4000 Hz

However…Keller and Valente (2006)…However…Keller and Valente (2006)…

N = >300 earsN = >300 ears

REDD as a function of frequency

0

5

10

15

Frequency (Hz)

R

Mean REDD

Next….with the same stroke of the Next….with the same stroke of the keyboard, the software will keyboard, the software will predictpredict the the LDL in dB HL from Pascoe (1988) and LDL in dB HL from Pascoe (1988) and add the add the averageaverage REDD to convert the REDD to convert the

Next….with the same stroke of the Next….with the same stroke of the keyboard, the software will keyboard, the software will predictpredict the the LDL in dB HL from Pascoe (1988) and LDL in dB HL from Pascoe (1988) and add the add the averageaverage REDD to convert the REDD to convert the add t eadd t e a e agea e age to co e t t eto co e t t e

LDL from dB HL to LDL in dB SPLLDL from dB HL to LDL in dB SPLadd t eadd t e a e agea e age to co e t t eto co e t t e

LDL from dB HL to LDL in dB SPLLDL from dB HL to LDL in dB SPL

8/6/2012

13

Frye 6500, 7000 and 8000Frye 6500, 7000 and 8000Frye 6500, 7000 and 8000Frye 6500, 7000 and 8000

REDDREDD LDLLDL

+

HL 250 500 1000 1500 2000 3000 4000 6000 8000Pascoe(1988)

0 116 109 106 109 112 113 110 110 111 97

5 118 111 108 111 114 115 112 112 113 99

10 118 111 108 111 114 115 112 112 113 99

15 117 110 107 110 113 114 111 111 112 98

20 116 109 106 109 112 113 110 110 111 97

25 120 113 110 113 116 117 114 114 115 101

30 121 114 111 114 117 118 115 115 116 102

35 120 113 110 113 116 117 114 114 115 101

40 122 115 112 115 118 119 116 116 117 103

45 124 117 114 117 120 121 118 118 119 105

50 126 119 116 119 122 123 120 120 121 107

55 127 120 117 120 123 124 121 121 122 108

1. Pascoe (1988)dB HL to LDL

(last column to right)

2. HL to SPL conversion

(ANSI S3.6-1989Table G-1) ineach cell)

==============250 1955 127 120 117 120 123 124 121 121 122 108

60 129 122 119 122 125 126 123 123 124 110

65 133 126 123 126 129 130 127 127 128 114

70 134 127 124 127 130 131 128 128 129 115

75 136 129 126 129 132 133 130 130 131 117

80 139 132 129 132 135 136 133 133 134 120

85 139 132 129 132 135 136 133 133 134 120

90 143 136 133 136 139 140 137 137 138 124

95 149 142 139 142 145 146 143 143 144 130

100 146 139 136 139 142 143 140 140 141 127

105 152 145 142 145 148 149 146 146 147 133

110 153 146 143 146 149 150 147 147 148 134

115 156 149 146 149 152 153 150 150 151 137

120 159 152 149 152 155 156 153 153 154 140

250 19500 121000 91500 122000 153000 15.54000 136000 138000 14==============

Predicted RESR90 (Dillon, 2012)

133 126 123 129 127 Frye Values

8/6/2012

14

Predicted LDL

Bentler and Bentler and Cooley (Cooley (2001)2001)Bentler and Bentler and Cooley (Cooley (2001)2001)

170170

500 Hz

y = 0.0821x + 108.17

R2 = 0.0255

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

170

0

10

20

30

40

50

60

70

80

90

10

0

11

0

12

0

Ind

ivid

ua

l L

DL

(in

dB

SP

L)

n=451

Mean LDL

1000 Hz

y = 0.1187x + 104.57

R2 = 0.0586

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

170

0

10

20

30

40

50

60

70

80

90

10

0

11

0

12

0

n=473

MeanLDL

Keller & Keller & ValenteValente((2006)2006)

N = 451 N = 473

4000 Hz

y = 0.1882x + 103.76

R2 = 0.1515

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

170

0

10

20

30

40

50

60

70

80

90

10

0

11

0

12

0

Hearing Threshold Level (in dB HL)

n=330

Mean LDL

Linear(n=330)

2000 Hz

y = 0.1946x + 105.81

R2 = 0.0959

0

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

160

170

0

10

20

30

40

50

60

70

80

90

10

0

11

0

12

0

Hearing Threshold Level (in dB HL)

Ind

ivid

ua

l L

DL

(in

dB

SP

L)

n=444

Mean LDLN = 444 N = 330

8/6/2012

15

How can one measure the individual REDD?

How can one measure the individual REDD?REDD?REDD?

Measuring Measuring REDDREDDMeasuring Measuring REDDREDD

Audiometer dial isAudiometer dial is70 dB HL70 dB HL

ProbeProbe

Insert Earphone fromAudiometer

Probe Microphone

8/6/2012

16

Frye 7000Frye 7000Frye 7000Frye 7000

Probe Microphone 82.1 dB

Frye 8000Frye 8000

VerifitVerifit

8/6/2012

17

Table used to document REDD in SPLTable used to document REDD in SPL(We’ll Come Back to This a Little Later) (We’ll Come Back to This a Little Later) Table used to document REDD in SPLTable used to document REDD in SPL(We’ll Come Back to This a Little Later) (We’ll Come Back to This a Little Later)

RT LT HL SPL REDD HL SPL REDD500 7070 7878 881000 2000 3000 4000

70707070

7070

7070

7070

78788282

8787

8080

7676

881212

1717

1010

66Add REDD to audio threshold to convert to dB SPL @ TM. Add REDD to audio threshold to convert to dB SPL @ TM.

Using the Dynamic Range of the Using the Dynamic Range of the Patient as thePatient as the TargetTarget

Using the Dynamic Range of the Using the Dynamic Range of the Patient as thePatient as the TargetTargetPatient as thePatient as the TargetTargetPatient as thePatient as the TargetTarget

SPLSPL--OO--GRAMGRAMSPLSPL--OO--GRAMGRAMLDL

Dynamic Range

SP

L

Threshold

dB

S

NAL-NL1 REAR Targets for50-65-80 dB SPL

8/6/2012

18

Press F2 from opening screen to open Real Press F2 from opening screen to open Real Ear Navigation ScreenEar Navigation Screen

Press F1 to enter audiogram and LDL in dB Press F1 to enter audiogram and LDL in dB ggHL.HL.

If individually measured LDL are not entered If individually measured LDL are not entered the software will the software will predictpredict LDL based on Pascoe LDL based on Pascoe

(1988)(1988)

Audio + Audio + measuredmeasured LDL of left ear in dB HLLDL of left ear in dB HL

8/6/2012

19

Hz dB ANSI- S3.6 HL 1989

PredictedThreshold (SPL)

My REDD Diff Measured ChangeThreshold SL of REAR(dB SPL)

250 10 19 29 13 -6 23 +6

500 20 12 32 15 +3 35 -3

1000 30 9 39 12 +3 42 -3

2000 40 15 55 16 +1 56 -1

3000 50 15.5 12 -3.5

4000 60 13 73 2 -11 62 -11

6000 70 13 10 -3

8000 80 13 93 12 -1 92 +1

Target screen beforecorrection

REARTarget

REIGTarget

Press F2 to enter realPress F2 to enter real--eareartarget screentarget screen

8/6/2012

20

REDD menu

[Menu] and scroll down to REDD[Menu] and scroll down to REDD

Hit → to obtain “Custom”Hit → to obtain “Custom”

REDD Custom

REDD via F2

8/6/2012

21

Corrected forEntered REDD

UncorrectedUncorrected

Final Final correctioncorrection

Notice REAR did not change, but SL changed between predicted and measured

8/6/2012

22